DATA CLUSTERING: APPLICATIONS IN ENGINEERING
|
|
- Laura Gaines
- 5 years ago
- Views:
Transcription
1 DATA CLUSTERING: APPLICATIONS IN ENGINEERING Zdravo Krpć Faculty of Electrcal Engneerng, Unversty of Ose Kneza Trpmra 2B, HR-3000 Ose E-mal: Goran Martnovć Faculty of Electrcal Engneerng, Unversty of Ose Kneza Trpmra 2B, HR-3000 Ose Phone: ; E-mal: goran.martnovc@etfos.hr Ivan Vazler Department of Mathematcs, Unversty of Ose Gaev trg 6, HR-3000 Ose Phone: ; E-mal: vazler@mathos.hr Abstract T n Dvdng a set S x ( x, x ) :, m dsunct subsets,, R (a set of vectors from a vector space n, m, such that S, 0,,,,, determnes a partton of the set S. The elements of such partton,, are called clusters. n R ) nto For practcal clusterng applcatons the number of all clusters s too bg and the problem of determnng the optmal partton n the least-squares sense s an NP-hard problem. In ths paper we wll consder some well-nown algorthms for searchng for an optmal LS-partton, lst some of the numerous applcatons of cluster analyss n engneerng and gve some practcal applcatons. Key words: data clusterng, engneerng, least squares. INTRODUCTION In short, clusterng problems are problems of dentfyng groups of ndvduals or obects that are smlar to each other but dfferent from those n other groups. Many web portals and nternet busnesses trac consumer 80
2 habts and tae advantage of these smlartes to target specfc offers to subgroups that are most lely to be receptve to them. Many search engnes cluster ther databases so they can offer smlar results (le boostores suggestng other boos by the same author, or boos wth smlar topcs, or boos from the same publsher, and so on). T n Dvdng a set S x ( x, x ) :, m dsunct subsets,, R (a set of vectors from a vector space n, m, such that S, 0,,,,, determnes a partton of the set S, whch wll be denoted by,, S n R ) nto. The elements of such partton,, are called clusters. The set of all parttons of the set S contanng clusters whch satsfy the propertes above wll be denoted by S, The number of all -parttons s.! S, ( ) and the goal of clusterng s to fnd the optmal partton n some sense. For practcal clusterng applcatons that number s too bg. The problem of clusterng can be dvded n several subproblems. Obects beng clustered need to be represented n a way that the clusterng algorthms can easly measure ther smlarty or dssmlarty (or dstance). Determnng the goal functon and an algorthm for clusterng (whch n most cases fnds only the approxmaton of the optmal partton) s another problem. Dependng on the algorthm, the problem of determnng the number of clusters can also arse. Clusterng algorthms can be classfed n several categores: - Herarchcal clusterng - Based on a tree model of data, t can be ether agglomeratve or dvsve. Agglomeratve clusterng begns wth each obect n ts own cluster and n each step clusters are oned based on smlarty. Its bad sde s that once two obects are n the same cluster, they stay n t tll the algorthm ends. Dvsve clusterng wors smlarly n the opposte drecton. m 8
3 - Parttonal clusterng - These are methods that teratvely mprove the parttonng by movng elements from one cluster to another, usually startng from a random partton. K-means and - medods are the most nown such algorthms. - Neural networ-based clusterng - Hgh dmensonal and large-scale data clusterng - These are methods based on reducng the dmensonalty of the problem. They nclude random samplng methods, densty-based methods and grd-based methods. If we defne a crtera functon on the set S, f : P F : then we can defne a partton S 0, f of all parttons of the set S contanng clusters by x S, 0, F f, x c (o) whch s optmal n the least-squares sense,.e. F ( o) mn S, F. The problem of determnng the optmal partton n the least-squares sense s an NP-hard problem. The -means algorthm can be used for searchng the optmal partton n the LS sense. 2 Algorthm : K-Means Input: Arbtrary -partton (0) T n of the set S x ( x, xn ) R :, m Output: Locally LS-optmal -partton t 0 (l) t) ( t ) ( t) ( t ) 2 calculate cluster centres c c,, c c 3 repeat 4 t t ( t ) t 5 ),, ( t ) ( ( x ( t ) ( t ) x ( t) ( t) such that x S : arg mn x c t) ( t ) ( t) ( t) 6 calculate new cluster centres c c,, c c ) 7 untl c t c (t ) 8 return ( ( t ) ( x ( t) ( t ) x It s easly shown that the -means algorthm monotonously reduces the crtera functon F. The algorthm often stops before reachng the optmal partton n the LS sense. The partton on whch the algorthm stops p p 82
4 depends on the choce of the ntal partton, and snce the algorthm s usually very quc, t s very common to run t multple tmes wth dfferent startng parttons to ncrease the chance of obtanng a better resultng partton. 2. APPLICATIONS 2.. Text clusterng There are many uses of clusterng n text analyss. Clusterng can be used to derve eywords, group artcles wth smlar topcs (ntally unnown), fnd possble synonyms n monolngual dctonares and n many other areas. To cluster textual data one must frst transform the textual data to a format that can be used n clusterng algorthms. One way to represent artcles would be to use ther references and represent ther connectons wth a graph. Ths representaton s sutable for clusterng wth herarchcal clusterng algorthms. Another way to pre-process textual data would be to represent t n the form of vectors (whch can be done n many ways dependng of our goal). The method of representaton used here s the vectorbased model descrbed n Berry (2004) and Srvastava and Saham (2009). The most common way of creatng a vector space model can be dvded n two stages. The frst stage s the extracton of content bearng terms (words or short phrases) and settng ther weght proportonal to the count of the correspondng term n the document. The second stage s to modfy the weghts so that the mportant terms get more emphass. The set of m documents would be represented by a set of vectors T n S x ( x, x ) R :, m where dmenson n of the vectors s equal to the number of n terms n the whole document collecton. The frst tas n stage one s to determne all the terms. Some terms n a document don t descrbe any mportant content (e.g. pronouns,...). Other terms may appear n all (or most) documents or only several documents. These words are usually fltered from the documents and do not appear n the vector representaton. In many languages some terms can be condensed to one due to conugaton or declenson. In the frst stage we create vectors f,, m of frequences of terms. The value of () f s set to the frequency of the th term n the th document. Note that the vectors are very sparse, snce many terms appear only n several documents. In the second stage, the term frequences are multpled by the nverse document frequency of a term n the document collecton. If we denote by W daglnm / w lnm / w n a dagonal matrx where w s the total number of documents contanng the th term, then the vector representaton of document s x' W f,, m. 83
5 Ths s done so that terms occurrng n almost all documents don t nfluence the clusterng results as much. 3.0 few occurences many occurences Fgure : Example of nverse document frequency weghts The last thng left to do s to normalze the vectors so that we observe the relatve frequency of terms. Normalzaton can be done n dfferent norms (,2, ) resultng n data on a correspondng n -dmensonal sphere. After the preprocessng step we have x x' x',, m. Although the -means algorthm s not very good for clusterng hgh dmensonal data the clusterng of normalzed vectors usng the -means should be done wth an extra step so that the centres of clusters also le on a unt sphere. If the Eucldan norm s used for normalzaton, a smple normalzaton of the centrods yelds the centres of clusters constraned on a unt sphere. Ths modfed algorthm s often called sphercal -means algorthm and s obtaned by addng the normalzaton step after steps 2 and 6 n Algorthm. ( t) c c ( t) Example. We tred out ths method of groupng of 0 short news artcles n Croatan found on the nternet. There were 290 dfferent terms. We dd not combne terms based on conugaton and declenson and therefore the results are not as good as they could be. Also, many of the artcles are too short to have enough groupng terms. Despte those shortcomngs, the results are satsfactory. These are the most frequent terms by clusters: ) utamca, ugovor, Zagreb, Masmr, postgao, gol, prva, navač,... 2) azna, sedala, prodaa, orsn, pad, automobl, vozla,... 3) porenuta, odvetn, bvš, uzeo, zbrsao, bane, mto,... 4) aptal, rebalans, povećat, cgareta, bane, pdv, porez, proračun, Terms n those clusters are characterstc terms for football, automobles, banng and poltcs. / c ( t) 84
6 2.2. Image analyss Cameras and other magng equpment are cheap, avalable and used n many areas. Wth that, the need for an automatc mage analyss has arsen. Often the goal s to dstngush smlar or dssmlar areas of an mage. In medcal scences clusterng can be appled to varous body scans for tumor dagnostcs. On satellte mages, clusterng can be used to dscern urban areas, felds, forests... Clusterng can also be used to fnd dfferently textured areas of an mage. To do any of these thngs, mages must often be pre-processed to show the dstngushng features. The most common mage attrbute used for mage clusterng s colour. Other smple mage characterstcs applcable for clusterng parameters are hue and saturaton. More complex mage propertes nclude pxel dstances and patterns. Some pattern recognton applcatons also requre large mage databases aganst whch canddate mages are compared. Example 2. In ths example -means clusterng method s used on a 256 colour greyscale mage, as proposed n Saha and Bandyopadhyay (2008). The mage s 256 pxels n wdth and heght. Dfferent areas of nterest are extracted from t based on the pxel colour value. Ths applcaton s common n satellte mage analyss, but t has some other uses, such as those descrbed n Tatrau and Mehta (2008). The number of clusters represent granularty of segments needed for the extracton from the mage. a) b) c) d) Fgure 2: Clusterng of a greyscale mage based on pxel colour value: a) Orgnal mage, b) =2, c) =3, d) =8. 85
7 Fgures 2 and 3 show two mages segmented nto dfferent numbers of clusters. As the number of cluster ncreases, more detaled mage analyss s done, but ncreasng the number of clusters beyond a certan threshold can mae clusterng meanngless. a) b) c) d) Fgure 3: -means clustered satellte mage: a). Orgnal mage, b) Separaton of heavy clouds wth =2, c) Separaton of lght and heavy clouds wth =4, d) Wth 8 clusters, there s no gan n enhancng cloud separaton. In order to dscover mportant areas of a greyscale mage we need to now how many clusters to loo for. There are many crtera for determnng that number, and many of them requre clusterng of the data for each. One of the easer ways s to use hstogram analyss of the mage. The procedure s as follows: Let G denote the set of all grey levels. For every shade of grey G we fnd ts frequency f, the number of pxels wth that colour. We detect the set of local maxmums n the mage hstogram S f f f f f, & We remove the local maxmums wth frequences below some emprcal threshold (for example f / 00 f max ). thr S 2, f S f f thr 86
8 From S 2 we remove the elements havng close peas (ther dfference n grey levels s below some emprcally determned threshold t ). S 3 S 2 f S, f S such that & t& f f, 2 2 The number S3 s the number of clusters to loo for, and the elements remanng n S 3 are good ntal centres for clusterng. a) b) Fgure 4: Grey level hstograms of mages from a) Fgure 2. b) Fgure 3. In Fgure 4 we can see that the frst mage has fluctuatng grey level frequences so we would use the frequency threshold f thr to remove low level peas. The grey level frequences of the second mage are more unform so we would use the threshold t to reduce the number of peas Applcaton n computer scence A possble applcaton of clusterng n computer scence can be n applcaton mappng n heterogeneous envronments. Advances of ths research can be found n Segel (2009). The heterogeneous computng envronment comprses of dfferent computers under dfferent loads. The goal s to fnd optmal groups of computers (computer clusters) whch are capable of performng varous applcaton tass. These systems can be found n varous computer cluster nstallatons, computer grds and cloud computng systems. Fndng the optmal computer(s) for solvng an applcaton-gven problem can often be NP hard, as there are many parameters whch descrbe each computer. Another dffculty arses due to the dfferent nature of these parameters (processor speed n MHz, RAM capacty n megabytes, networ throughput n megabts per second, etc.). Ths mples that normalzaton of parameters has to be done frst. After that, snce the mappng system uses dfferent preferences on dfferent parameters for every applcaton tas, analyss has to be performed whch evaluates parameter mpact on canddate sutablty. Ths s done by usng weghts, and multplyng normalzed parameter values wth them. The metrc used for measurng the dstance and for calculatng the smlarty matrx depends on a number of parameters for each mappng canddate. In Table, 87
9 ten parameters whch descrbe mappng canddates are presented. The range of values (Mnmum value and Maxmum value) used durng calculaton s also gven for each canddate. After normalzaton, statc and dynamc data are combned together to form current computer mappng canddate (MC) state. Table : Canddate parameters Parameter Measurng unt Mnmum value Maxmum value Processor speed MHz memory capacty MB hard ds capacty GB networ throughput Mbt/s 000 Operatng system -3 3 Processor load % 0 00 Memory load % 0 00 Ds space usage % 0 00 Networ traffc % 0 00 St at c Dynamc Example 3. For the purpose of vsualzaton smplcty, greatest weghts were gven to the frst two parameters (avalable CPU speed and avalable RAM memory), meanng that these are mostly requred by the applcaton tas. Runnng the mappng n ths envronment, wth -means clusterng ( 4 ) gves selectons shown on Fgure 5. Fgure 5: Results of a mappng system, usng -means clusterng method It s obvous that only computer cluster has the benefts from both parameters, formng the most powerful computer cluster. Computer cluster 2 comprses computers wth great processng power, but they lac RAM memory. However, ths cluster has ts uses. Many applcatons depend almost exclusvely on processor 88
10 speed. The thrd computer cluster holds most of the computers, whch have lower performance. Ths cluster contans unwanted canddates, whch are ether heavly loaded already, or ther hardware s nsuffcent. Last computer cluster holds computers wth large amount of RAM memory. These are the approprate canddates for applcaton tass whch are hungry memory-wse. In concluson, there are many dfferent applcatons of clusterng. They dffer n data representaton, goal functons or method of clusterng and that s the reason behnd the ncreasng number of artcles n ths feld. REFERENCES Bandyopadhyay, S. and Saha, S. (2008), Unsupervsed pxel classfcaton n satellte magery usng a new multobectve symmetry based clusterng approach, TENCON, Inda, pp. -6. Berry, M. W. (2004), Survey of text mnng: Clusterng, classfcaton, and retreval, Sprnger, Berln Dubes, R.C. and Jan, A. K. (988), Algorthms for clusterng data, Prentce Hall, New Jersey Evertt, B. S., Landau, S. and Leese, M. (200), Cluster analyss, Wley, London Fran, E. and Wtten, I. H. (2005), Data mnng: Practcal machne learnng tools and technques, Morgan Kaufmann Gan, G., Ma, C. and Wu, J. (2007), Data clusterng: theory, algorthms, and applcatons, SIAM, Phladelpha Han, J. and Kamber, M. (2006), Data mnng: concepts and technques, Morgan Kaufmann Hartgan, J. A. (975) Clusterng algorthms, Wley Jauga, K., Soolows, A. and Boc, H. H. (2002), Classfcaton, clusterng and data analyss, Sprnger, Berln Jng, T., Oscar, C. A., Ruobng, Z., Weyu, Y. and Zhdng, Y. (2008), An adaptve unsupervsed approach toward pxel clusterng and color mage segmentaton, Elsever Kaufman, L. and Rousseeuw, P. J. (2005), Fndng groups n data: an ntroducton to cluster analyss, Jonh Wley & Sons, Hoboen Kogan, J. (2007), Introducton to clusterng large and hgh-dmensonal data, Cambrdge Unversty Press Mehta, A. and Tatrau, S. (2008), Image segmentaton usng -means clusterng, EM and Normalzed Cuts, Department of EECS report, Unversty Of Calforna Segel H. J. (2009), Stochastcally robust resource management n heterogeneous parallel computng systems, ISPAN, USA, pp. -2. Srvastava, A. and Saham, M. (2009), Text mnng: Classfcaton, clusterng, and applcatons, Chapman & Hall 89
A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION
1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute
More informationOutline. Type of Machine Learning. Examples of Application. Unsupervised Learning
Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton
More informationA Binarization Algorithm specialized on Document Images and Photos
A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a
More informationSubspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;
Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features
More informationFEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur
FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents
More informationCluster Analysis of Electrical Behavior
Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School
More informationMachine Learning: Algorithms and Applications
14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of
More informationCourse Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms
Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques
More informationUnsupervised Learning
Pattern Recognton Lecture 8 Outlne Introducton Unsupervsed Learnng Parametrc VS Non-Parametrc Approach Mxture of Denstes Maxmum-Lkelhood Estmates Clusterng Prof. Danel Yeung School of Computer Scence and
More informationContent Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers
IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth
More informationLearning the Kernel Parameters in Kernel Minimum Distance Classifier
Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department
More informationCS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 15
CS434a/541a: Pattern Recognton Prof. Olga Veksler Lecture 15 Today New Topc: Unsupervsed Learnng Supervsed vs. unsupervsed learnng Unsupervsed learnng Net Tme: parametrc unsupervsed learnng Today: nonparametrc
More informationHierarchical clustering for gene expression data analysis
Herarchcal clusterng for gene expresson data analyss Gorgo Valentn e-mal: valentn@ds.unm.t Clusterng of Mcroarray Data. Clusterng of gene expresson profles (rows) => dscovery of co-regulated and functonally
More informationUnsupervised Learning and Clustering
Unsupervsed Learnng and Clusterng Why consder unlabeled samples?. Collectng and labelng large set of samples s costly Gettng recorded speech s free, labelng s tme consumng 2. Classfer could be desgned
More informationA Fast Content-Based Multimedia Retrieval Technique Using Compressed Data
A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,
More informationMULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION
MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and
More informationFeature Reduction and Selection
Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components
More informationMachine Learning. Topic 6: Clustering
Machne Learnng Topc 6: lusterng lusterng Groupng data nto (hopefully useful) sets. Thngs on the left Thngs on the rght Applcatons of lusterng Hypothess Generaton lusters mght suggest natural groups. Hypothess
More informationA new segmentation algorithm for medical volume image based on K-means clustering
Avalable onlne www.jocpr.com Journal of Chemcal and harmaceutcal Research, 2013, 5(12):113-117 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCRC5 A new segmentaton algorthm for medcal volume mage based
More informationSLAM Summer School 2006 Practical 2: SLAM using Monocular Vision
SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,
More informationUser Authentication Based On Behavioral Mouse Dynamics Biometrics
User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA
More informationParallelism for Nested Loops with Non-uniform and Flow Dependences
Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr
More informationNUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS
ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana
More informationS1 Note. Basis functions.
S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type
More informationThe Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique
//00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy
More informationUnsupervised Learning and Clustering
Unsupervsed Learnng and Clusterng Supervsed vs. Unsupervsed Learnng Up to now we consdered supervsed learnng scenaro, where we are gven 1. samples 1,, n 2. class labels for all samples 1,, n Ths s also
More informationA Clustering Algorithm for Key Frame Extraction Based on Density Peak
Journal of Computer and Communcatons, 2018, 6, 118-128 http://www.scrp.org/ournal/cc ISSN Onlne: 2327-5227 ISSN Prnt: 2327-5219 A Clusterng Algorthm for Key Frame Extracton Based on Densty Peak Hong Zhao
More informationClustering. A. Bellaachia Page: 1
Clusterng. Obectves.. Clusterng.... Defntons... General Applcatons.3. What s a good clusterng?. 3.4. Requrements 3 3. Data Structures 4 4. Smlarty Measures. 4 4.. Standardze data.. 5 4.. Bnary varables..
More informationOptimizing Document Scoring for Query Retrieval
Optmzng Document Scorng for Query Retreval Brent Ellwen baellwe@cs.stanford.edu Abstract The goal of ths project was to automate the process of tunng a document query engne. Specfcally, I used machne learnng
More informationObject-Based Techniques for Image Retrieval
54 Zhang, Gao, & Luo Chapter VII Object-Based Technques for Image Retreval Y. J. Zhang, Tsnghua Unversty, Chna Y. Y. Gao, Tsnghua Unversty, Chna Y. Luo, Tsnghua Unversty, Chna ABSTRACT To overcome the
More informationLecture 5: Multilayer Perceptrons
Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented
More informationCS 534: Computer Vision Model Fitting
CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust
More informationImage Representation & Visualization Basic Imaging Algorithms Shape Representation and Analysis. outline
mage Vsualzaton mage Vsualzaton mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and Analyss outlne mage Representaton & Vsualzaton Basc magng Algorthms Shape Representaton and
More informationUB at GeoCLEF Department of Geography Abstract
UB at GeoCLEF 2006 Mguel E. Ruz (1), Stuart Shapro (2), June Abbas (1), Slva B. Southwck (1) and Davd Mark (3) State Unversty of New York at Buffalo (1) Department of Lbrary and Informaton Studes (2) Department
More informationSupport Vector Machines
/9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.
More informationMaximum Variance Combined with Adaptive Genetic Algorithm for Infrared Image Segmentation
Internatonal Conference on Logstcs Engneerng, Management and Computer Scence (LEMCS 5) Maxmum Varance Combned wth Adaptve Genetc Algorthm for Infrared Image Segmentaton Huxuan Fu College of Automaton Harbn
More informationEdge Detection in Noisy Images Using the Support Vector Machines
Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona
More informationLoad-Balanced Anycast Routing
Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance
More informationDecision Strategies for Rating Objects in Knowledge-Shared Research Networks
Decson Strateges for Ratng Objects n Knowledge-Shared Research etwors ALEXADRA GRACHAROVA *, HAS-JOACHM ER **, HASSA OUR ELD ** OM SUUROE ***, HARR ARAKSE *** * nsttute of Control and System Research,
More informationClassifier Selection Based on Data Complexity Measures *
Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.
More informationHierarchical agglomerative. Cluster Analysis. Christine Siedle Clustering 1
Herarchcal agglomeratve Cluster Analyss Chrstne Sedle 19-3-2004 Clusterng 1 Classfcaton Basc (unconscous & conscous) human strategy to reduce complexty Always based Cluster analyss to fnd or confrm types
More informationAn Image Fusion Approach Based on Segmentation Region
Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua
More informationMachine Learning 9. week
Machne Learnng 9. week Mappng Concept Radal Bass Functons (RBF) RBF Networks 1 Mappng It s probably the best scenaro for the classfcaton of two dataset s to separate them lnearly. As you see n the below
More informationSteps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices
Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between
More informationTerm Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task
Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto
More informationPictures at an Exhibition
1 Pctures at an Exhbton Stephane Kwan and Karen Zhu Department of Electrcal Engneerng Stanford Unversty, Stanford, CA 9405 Emal: {skwan1, kyzhu}@stanford.edu Abstract An mage processng algorthm s desgned
More informationKeyword-based Document Clustering
Keyword-based ocument lusterng Seung-Shk Kang School of omputer Scence Kookmn Unversty & AIrc hungnung-dong Songbuk-gu Seoul 36-72 Korea sskang@kookmn.ac.kr Abstract ocument clusterng s an aggregaton of
More informationRelated-Mode Attacks on CTR Encryption Mode
Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory
More informationQuery Clustering Using a Hybrid Query Similarity Measure
Query clusterng usng a hybrd query smlarty measure Fu. L., Goh, D.H., & Foo, S. (2004). WSEAS Transacton on Computers, 3(3), 700-705. Query Clusterng Usng a Hybrd Query Smlarty Measure Ln Fu, Don Hoe-Lan
More information12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification
Introducton to Artfcal Intellgence V22.0472-001 Fall 2009 Lecture 24: Nearest-Neghbors & Support Vector Machnes Rob Fergus Dept of Computer Scence, Courant Insttute, NYU Sldes from Danel Yeung, John DeNero
More informationWishing you all a Total Quality New Year!
Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma
More informationProblem Definitions and Evaluation Criteria for Computational Expensive Optimization
Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty
More informationA mathematical programming approach to the analysis, design and scheduling of offshore oilfields
17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and
More informationOn Some Entertaining Applications of the Concept of Set in Computer Science Course
On Some Entertanng Applcatons of the Concept of Set n Computer Scence Course Krasmr Yordzhev *, Hrstna Kostadnova ** * Assocate Professor Krasmr Yordzhev, Ph.D., Faculty of Mathematcs and Natural Scences,
More information1. Introduction. Abstract
Image Retreval Usng a Herarchy of Clusters Danela Stan & Ishwar K. Seth Intellgent Informaton Engneerng Laboratory, Department of Computer Scence & Engneerng, Oaland Unversty, Rochester, Mchgan 48309-4478
More informationAn Improved Image Segmentation Algorithm Based on the Otsu Method
3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,
More informationFuzzy C-Means Initialized by Fixed Threshold Clustering for Improving Image Retrieval
Fuzzy -Means Intalzed by Fxed Threshold lusterng for Improvng Image Retreval NAWARA HANSIRI, SIRIPORN SUPRATID,HOM KIMPAN 3 Faculty of Informaton Technology Rangst Unversty Muang-Ake, Paholyotn Road, Patumtan,
More informationRange images. Range image registration. Examples of sampling patterns. Range images and range surfaces
Range mages For many structured lght scanners, the range data forms a hghly regular pattern known as a range mage. he samplng pattern s determned by the specfc scanner. Range mage regstraton 1 Examples
More informationHistogram based Evolutionary Dynamic Image Segmentation
Hstogram based Evolutonary Dynamc Image Segmentaton Amya Halder Computer Scence & Engneerng Department St. Thomas College of Engneerng & Technology Kolkata, Inda amya_halder@ndatmes.com Arndam Kar and
More informationAn Optimal Algorithm for Prufer Codes *
J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,
More informationA Simple Methodology for Database Clustering. Hao Tang 12 Guangdong University of Technology, Guangdong, , China
for Database Clusterng Guangdong Unversty of Technology, Guangdong, 0503, Chna E-mal: 6085@qq.com Me Zhang Guangdong Unversty of Technology, Guangdong, 0503, Chna E-mal:64605455@qq.com Database clusterng
More informationSum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints
Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan
More informationLobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide
Lobachevsky State Unversty of Nzhn Novgorod Polyhedron Quck Start Gude Nzhn Novgorod 2016 Contents Specfcaton of Polyhedron software... 3 Theoretcal background... 4 1. Interface of Polyhedron... 6 1.1.
More informationType-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data
Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES
More informationOutline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:
Self-Organzng Maps (SOM) Turgay İBRİKÇİ, PhD. Outlne Introducton Structures of SOM SOM Archtecture Neghborhoods SOM Algorthm Examples Summary 1 2 Unsupervsed Hebban Learnng US Hebban Learnng, Cntd 3 A
More informationWeb Mining: Clustering Web Documents A Preliminary Review
Web Mnng: Clusterng Web Documents A Prelmnary Revew Khaled M. Hammouda Department of Systems Desgn Engneerng Unversty of Waterloo Waterloo, Ontaro, Canada 2L 3G1 hammouda@pam.uwaterloo.ca February 26,
More informationA Two-Stage Algorithm for Data Clustering
A Two-Stage Algorthm for Data Clusterng Abdolreza Hatamlou 1 and Salwan Abdullah 2 1 Islamc Azad Unversty, Khoy Branch, Iran 2 Data Mnng and Optmsaton Research Group, Center for Artfcal Intellgence Technology,
More informationSupport Vector Machines
Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned
More informationA Fast Visual Tracking Algorithm Based on Circle Pixels Matching
A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng
More informationKent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming
CS 4/560 Desgn and Analyss of Algorthms Kent State Unversty Dept. of Math & Computer Scence LECT-6 Dynamc Programmng 2 Dynamc Programmng Dynamc Programmng, lke the dvde-and-conquer method, solves problems
More informationFace Recognition Method Based on Within-class Clustering SVM
Face Recognton Method Based on Wthn-class Clusterng SVM Yan Wu, Xao Yao and Yng Xa Department of Computer Scence and Engneerng Tong Unversty Shangha, Chna Abstract - A face recognton method based on Wthn-class
More informationModular PCA Face Recognition Based on Weighted Average
odern Appled Scence odular PCA Face Recognton Based on Weghted Average Chengmao Han (Correspondng author) Department of athematcs, Lny Normal Unversty Lny 76005, Chna E-mal: hanchengmao@163.com Abstract
More informationDesign of Structure Optimization with APDL
Desgn of Structure Optmzaton wth APDL Yanyun School of Cvl Engneerng and Archtecture, East Chna Jaotong Unversty Nanchang 330013 Chna Abstract In ths paper, the desgn process of structure optmzaton wth
More informationAn Efficient Genetic Algorithm with Fuzzy c-means Clustering for Traveling Salesman Problem
An Effcent Genetc Algorthm wth Fuzzy c-means Clusterng for Travelng Salesman Problem Jong-Won Yoon and Sung-Bae Cho Dept. of Computer Scence Yonse Unversty Seoul, Korea jwyoon@sclab.yonse.ac.r, sbcho@cs.yonse.ac.r
More information6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour
6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the
More informationA Clustering Algorithm for Chinese Adjectives and Nouns 1
Clusterng lgorthm for Chnese dectves and ouns Yang Wen, Chunfa Yuan, Changnng Huang 2 State Key aboratory of Intellgent Technology and System Deptartment of Computer Scence & Technology, Tsnghua Unversty,
More informationDetection of an Object by using Principal Component Analysis
Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,
More informationA Deflected Grid-based Algorithm for Clustering Analysis
A Deflected Grd-based Algorthm for Clusterng Analyss NANCY P. LIN, CHUNG-I CHANG, HAO-EN CHUEH, HUNG-JEN CHEN, WEI-HUA HAO Department of Computer Scence and Informaton Engneerng Tamkang Unversty 5 Yng-chuan
More informationA New Feature of Uniformity of Image Texture Directions Coinciding with the Human Eyes Perception 1
A New Feature of Unformty of Image Texture Drectons Concdng wth the Human Eyes Percepton Xng-Jan He, De-Shuang Huang, Yue Zhang, Tat-Mng Lo 2, and Mchael R. Lyu 3 Intellgent Computng Lab, Insttute of Intellgent
More informationElectrical analysis of light-weight, triangular weave reflector antennas
Electrcal analyss of lght-weght, trangular weave reflector antennas Knud Pontoppdan TICRA Laederstraede 34 DK-121 Copenhagen K Denmark Emal: kp@tcra.com INTRODUCTION The new lght-weght reflector antenna
More informationAn Internal Clustering Validation Index for Boolean Data
BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 16, No 6 Specal ssue wth selecton of extended papers from 6th Internatonal Conference on Logstc, Informatcs and Servce Scence
More informationVirtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory
Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process
More informationSurvey of Cluster Analysis and its Various Aspects
Harmnder Kaur et al, Internatonal Journal of Computer Scence and Moble Computng, Vol.4 Issue.0, October- 05, pg. 353-363 Avalable Onlne at www.csmc.com Internatonal Journal of Computer Scence and Moble
More informationReducing Frame Rate for Object Tracking
Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg
More informationSmoothing Spline ANOVA for variable screening
Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory
More informationLearning-Based Top-N Selection Query Evaluation over Relational Databases
Learnng-Based Top-N Selecton Query Evaluaton over Relatonal Databases Lang Zhu *, Wey Meng ** * School of Mathematcs and Computer Scence, Hebe Unversty, Baodng, Hebe 071002, Chna, zhu@mal.hbu.edu.cn **
More informationK-means and Hierarchical Clustering
Note to other teachers and users of these sldes. Andrew would be delghted f you found ths source materal useful n gvng your own lectures. Feel free to use these sldes verbatm, or to modfy them to ft your
More informationAccessibility Analysis for the Automatic Contact and Non-contact Inspection on Coordinate Measuring Machines
Proceedngs of the World Congress on Engneerng 008 Vol I Accessblty Analyss for the Automatc Contact and Non-contact Inspecton on Coordnate Measurng Machnes B. J. Álvarez, P. Fernández, J. C. Rco and G.
More informationGraph-based Clustering
Graphbased Clusterng Transform the data nto a graph representaton ertces are the data ponts to be clustered Edges are eghted based on smlarty beteen data ponts Graph parttonng Þ Each connected component
More informationAvailable online at Available online at Advanced in Control Engineering and Information Science
Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced
More informationImprovement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration
Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,
More informationFitting: Deformable contours April 26 th, 2018
4/6/08 Fttng: Deformable contours Aprl 6 th, 08 Yong Jae Lee UC Davs Recap so far: Groupng and Fttng Goal: move from array of pxel values (or flter outputs) to a collecton of regons, objects, and shapes.
More informationClassifying Acoustic Transient Signals Using Artificial Intelligence
Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)
More informationData Mining: Model Evaluation
Data Mnng: Model Evaluaton Aprl 16, 2013 1 Issues: Evaluatng Classfcaton Methods Accurac classfer accurac: predctng class label predctor accurac: guessng value of predcted attrbutes Speed tme to construct
More informationKOHONEN'S SELF ORGANIZING NETWORKS WITH "CONSCIENCE"
Kohonen's Self Organzng Maps and ther use n Interpretaton, Dr. M. Turhan (Tury) Taner, Rock Sold Images Page: 1 KOHONEN'S SELF ORGANIZING NETWORKS WITH "CONSCIENCE" By: Dr. M. Turhan (Tury) Taner, Rock
More informationTECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.
TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of
More informationA CALCULATION METHOD OF DEEP WEB ENTITIES RECOGNITION
A CALCULATION METHOD OF DEEP WEB ENTITIES RECOGNITION 1 FENG YONG, DANG XIAO-WAN, 3 XU HONG-YAN School of Informaton, Laonng Unversty, Shenyang Laonng E-mal: 1 fyxuhy@163.com, dangxaowan@163.com, 3 xuhongyan_lndx@163.com
More informationLECTURE : MANIFOLD LEARNING
LECTURE : MANIFOLD LEARNING Rta Osadchy Some sldes are due to L.Saul, V. C. Raykar, N. Verma Topcs PCA MDS IsoMap LLE EgenMaps Done! Dmensonalty Reducton Data representaton Inputs are real-valued vectors
More informationA Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines
A Modfed Medan Flter for the Removal of Impulse Nose Based on the Support Vector Machnes H. GOMEZ-MORENO, S. MALDONADO-BASCON, F. LOPEZ-FERRERAS, M. UTRILLA- MANSO AND P. GIL-JIMENEZ Departamento de Teoría
More informationSkew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach
Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research
More information